EP1286511B1 - Peak power control in a Multi-tone transmission system - Google Patents
Peak power control in a Multi-tone transmission system Download PDFInfo
- Publication number
- EP1286511B1 EP1286511B1 EP02255456A EP02255456A EP1286511B1 EP 1286511 B1 EP1286511 B1 EP 1286511B1 EP 02255456 A EP02255456 A EP 02255456A EP 02255456 A EP02255456 A EP 02255456A EP 1286511 B1 EP1286511 B1 EP 1286511B1
- Authority
- EP
- European Patent Office
- Prior art keywords
- symbol
- data stream
- analogue front
- symbol data
- analogue
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Expired - Lifetime
Links
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L27/00—Modulated-carrier systems
- H04L27/0002—Modulated-carrier systems analog front ends; means for connecting modulators, demodulators or transceivers to a transmission line
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L27/00—Modulated-carrier systems
- H04L27/26—Systems using multi-frequency codes
- H04L27/2601—Multicarrier modulation systems
- H04L27/2614—Peak power aspects
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L27/00—Modulated-carrier systems
- H04L27/32—Carrier systems characterised by combinations of two or more of the types covered by groups H04L27/02, H04L27/10, H04L27/18 or H04L27/26
- H04L27/34—Amplitude- and phase-modulated carrier systems, e.g. quadrature-amplitude modulated carrier systems
- H04L27/36—Modulator circuits; Transmitter circuits
- H04L27/366—Arrangements for compensating undesirable properties of the transmission path between the modulator and the demodulator
- H04L27/367—Arrangements for compensating undesirable properties of the transmission path between the modulator and the demodulator using predistortion
- H04L27/368—Arrangements for compensating undesirable properties of the transmission path between the modulator and the demodulator using predistortion adaptive predistortion
Definitions
- a system using multiple tone signalling generally uses the Fourier Transform and its inverse to convert the information between time and frequency domains.
- Two examples of this type of modulation scheme are: (a) DMT (Discrete Multi-Tone) as used in systems such as ADSL (Asymmetric Digital Subscriber Loop); and (b) COFDM (Carrierless Orthogonal Frequency Division Multiplex), a standard widely adopted for digital terrestrial TV broadcasting.
- DMT Discrete Multi-Tone
- ADSL Asymmetric Digital Subscriber Loop
- COFDM Carrierless Orthogonal Frequency Division Multiplex
- the data to be transmitted are sub-divided (multiplexed) across a number of distinct frequencies (sometimes also referred to as tones or sub-carriers) which are all integer multiples of a fixed basic frequency.
- the individual tones making up the group are spaced apart by this basic frequency.
- the number of tones used in different systems and within an individual system can vary, anywhere from 10 or so; e.g. for a low bandwidth ADSL upstream link, up to several thousand, e.g. an "8K-carrier" COFDM digital TV transmission.
- the key algorithm common to the communication systems under consideration is the Fourier Transform, a mathematical scheme in which a time-varying signal is represented not as a set of values in time but as the sum of a set of sinusoidal waveforms. Each sinusoid in the set has a distinct frequency which is an integer multiple of a base frequency called the analysis frequency.
- Fourier Transform theory shows that any varying signal can be alternately represented in this way, by defining the unique set of amplitude and phase values for the individual sinusoids which sum together to form the signal wave-shape.
- the size of the set of sinusoids is infinite and the spacing of the individual frequencies is infinitesimal.
- DFT Discrete Fourier Transform
- the term 'discrete' is used because the data is processed as a set of distinct samples, not a continuous signal. When a finite sequence of samples is transformed in this way, the size of the set of sinusoids that represent the signal in the frequency domain is also finite.
- the term 'discrete' should be assumed.
- the normal (“forward") Fourier Transform is used to convert from a series of samples taken in the time domain into an equivalent representation of the same information, namely as a series of values in the frequency domain, describing the amplitude and phase of each of a set of harmonically related sinusoidal waveforms.
- the reverse process, the Inverse Fourier Transform performs the opposite operation, summing the waveforms described by the individual amplitude and phase values to re-create a composite waveform as a series of samples in the time domain.
- the Fourier Transform and its inverse are relatively complex functions, but they may be implemented without difficulty using well-known algorithms on a digital signal processor.
- highly efficient versions of the transforms are known, commonly called the Fast Fourier Transform (FFT) and the Inverse FFT or IFFT, which operate on sample sequences whose lengths are powers of 2, e.g. 256 points or 512 points.
- FFT Fast Fourier Transform
- IFFT Inverse FFT
- the FFT and IFFT together provide for efficient encoding and decoding of signals.
- a set of data bits may be encoded by the IFFT, choosing particular combinations of amplitude and phase for each of the constituent frequency components to represent different data values.
- the IFFT is performed to create a time-domain signal which is then transmitted.
- the (forward) FFT is used at the receiver to reverse the process.
- the FFT is applied to the set of samples making up each received symbol, to reconstruct values of amplitude and phase for each of the tones in use.
- the values obtained by this process are not exactly the same as were initially encoded, for various reasons, including particularly the presence of noise introduced along the transmission path of the signal. Noise is unavoidable in any practical system.
- the original data may be recovered with an acceptable level of reliability, provided the system has been configured appropriately, taking into account the signal-transfer characteristics of the transmission path.
- cyclic prefix time in which what is transmitted is a portion of the signal extracted from the end of the immediately following symbol.
- cyclic prefix time derives from the fact that the short sequence has been used as a prefix to the new symbol and is cyclically congruent with it. Note that after equalization, the signal received during the cyclic prefix time is ignored by the receiver. In COFDM, the delay period is called the “guard time”; no signal is transmitted during this time.
- the IFFT-FFT (encoding-decoding) process provides for great flexibility in the communications system. Different frequencies in the spectrum covered by the set of tones may have different characteristics in respect of noisiness and attenuation over the communication link (e.g. the phone line in the case of an ADSL system). By varying the encoding details tone by tone, this may be accounted for, so as to maximize the number of bits carried by the symbol in total, even when a particular single tone can only carry a small number of bits.
- U.S. Patent No. 4,679,227 which describes multi-tone encoding schemes, presents techniques for accomplishing this.
- the waveform resulting from the IFFT can in principle have very large peak values in it - relative to the average amplitude of the signal as a whole - at points where the particular phases of the individual tones happen to sum together in the same direction. For example, if all tones were using encoded simple 2-bit quadrature modulation, and all the data bits being modulated were zero (or more generally if the same pair of bit values were modulating each tone), then at the start of the time domain symbol created by the IFFT there would be a high amplitude "spike", since each component waveform would have a real positive value 0.707 times its peak amplitude, and these would all sum together in the same direction. By contrast, if there is a general haphazard distribution of 1's and 0's in the data, the expected peak value in the average symbol would be much lower, although once in a while peaks will still occur.
- the signal On observing the output from a sequence of IFFT operations used to encode a (generalised) data sequence for transmission, the signal is seen to have a sample amplitude distribution which is very like random noise, when considered on a statistical basis.
- the most frequently occurring sample amplitudes are those near zero (the central point - the distribution is symmetrical either side of zero). Higher amplitudes are less likely, but still occur, and there is a gradual reduction in likelihood of occurrence with increasing amplitude.
- the overall statistical properties of the sequence are complex.
- one simple measure of the properties of signals generally is their crest factor.
- the crest factor of a repetitive signal is defined as the ratio of its peak amplitude to its average (RMS) amplitude.
- RMS average
- Different types of waveforms can have very different crest factors, depending on their shape. For example a simple pulse waveform, where the signal jumps between just two levels +A and -A, has a crest factor of 1, i.e. the average and peak levels of the signal are the same.
- a simple continuous sine wave has a crest factor of 2 (1.4142135).
- Other wave shapes can be envisaged having widely differing crest factors.
- the definition of crest factor is adjusted. This is necessary, in order to take into account the statistical spread of amplitude values.
- the effective crest factor to be the ratio of a threshold level to the average (RMS) level of the signal overall, where the threshold level is that which only some particular small fraction (e.g. 1/10,000,000 th , or 10 -7 ) of the generated samples will equal or exceed.
- any regular patterns in it may be broken up.
- the distribution of the data bit values going forward into the encoder becomes more haphazard, and so the likelihood of coherence between the phases of the different tones is drastically reduced. This diminishes the frequency with which spikes appear in the time-domain signal, even for a completely regular input stream (e.g. all 1s), relative to that which would apply without scrambling.
- a completely regular input stream e.g. all 1s
- IFFT-based encoding One major problem with IFFT-based encoding, so far as the design of any practical system is concerned, is that the time domain signal created has characteristics which make it more difficult and/or more expensive to carry through the later stages of the transmission path. For example, the bandwidth of the signal may in some cases be as wide as can theoretically be carried by the discrete sample sequence. Any subsequent processing of the signal, post-IFFT, must therefore be carefully designed to minimise distortions of the signal caused by frequency-dependent variations (e.g. in gain or phase-shift), which are typically worst at the highest frequencies.
- frequency-dependent variations e.g. in gain or phase-shift
- the first problem is that the dynamic range of the digital-to-analogue converter (DAC) must be large, requiring a relatively high number of bits of resolution (typically between 14 and 16 for ADSL). This makes the DAC hard to design, especially since it is running at high sampling rates (in the order of 1-2 MHz or higher for ADSL, and higher still for COFDM).
- the input circuitry In a receiver for the transmitted signal, the input circuitry must also have a high dynamic range and low noise and distortion; equally its analogue-to-digital converter must have high linearity and resolution.
- the second aspect which is usually considered even more serious, is that it is extremely difficult to design the amplification stages of the transmitter to both yield the high linearity which is needed and also maintain good power efficiency.
- the amplifier also called the "line-driver" in the case of ADSL
- the amplifier must be able to handle signal peaks several times higher than the average signal level on the line, it becomes necessary to run its power supply at a far higher voltage than the average signal level would require, if the signal's crest factor were lower.
- Typical power efficiencies for amplifiers in present-day ADSL system designs are therefore significantly lower than in some other types of transmission system e.g. 15-20% as against 40% or more.
- WO99/18662 to Ericsson describes one approach to minimise effects of peaks in transmitted power in a multi-carrier DSL-type transmission system.
- an amplifier circuit arranged for driving the line from an analogue input has two power supplies, of higher and lower voltage.
- a controller causes power to be supplied from the lower voltage power supply when the magnitude of the input signal is less than a threshold, and from the higher voltage power supply when the magnitude of the input signal is higher than the threshold.
- EP 1 104 140 discloses a multi-tone signalling transmitter with an alternative solution for reducing the crest factor.
- the modelling of the amplifier is used to configure sign-switches in the transmitter.
- the invention solves the above mentioned problems by a method as claimed in independent claim 1, a device as claimed in independent claim 5 and a system as claimed in independent claim 9.
- a modulation method for multiple-tone signalling using a system with an analogue front end comprising the steps of: feeding a symbol data stream of multiple tone symbols to an analogue front end and to a model; in the model, modelling the peak amplitude that will be present in the symbol data stream after subsequent processing by the analogue front end; feeding forward a control signal based on the modelled peak amplitude from the model to the analogue front end; and outputting the symbol data stream through the analogue front end under the control of the control signal.
- the analogue front end includes an amplifier such as a line driver operable from a plurality of different voltage levels and the control signal selects one of the plurality of different voltage levels in the line driver.
- an amplifier such as a line driver operable from a plurality of different voltage levels and the control signal selects one of the plurality of different voltage levels in the line driver.
- maximum power efficiency can be obtained.
- a lower voltage power supply is used, to provide maximum power efficiency.
- the maximum signal amplitude which can be amplified linearly is limited, and an input level that is too high will cause unwanted clipping of the signal in the amplifier.
- use of the higher voltage supply will allow higher signal amplitudes to be dealt with correctly, but will incur higher power consumption and hence lower efficiency.
- the line driver or other amplifier can be switched to use a higher voltage supply when a larger signal amplitude will be received.
- Amplifiers cannot switch power instantly; some delay is incurred before the new supply is fully connected and available.
- predicting the power in the symbol in advance it is possible to make the change-over at a convenient time when the delay will not cause significant corruption of the transmitted signal.
- system can also compensate in the model for any preprocessing provided in the analogue front end.
- the control signal need not be used solely to control an amplifier.
- the signal may also be used to influence preprocessing of the symbol data stream, or even the digital to analogue conversion.
- the method may also applied in a system which processes an input data stream through a plurality of intermediate processing stages and corresponding stages of intermediate data to generate the symbol data stream. If the modelled peak amplitude in a particular symbol in the symbol data stream exceeds a predetermined threshold, the intermediate data can be amended such that the input data is still represented by the intermediate data, and the subsequent intermediate processing stages on the intermediate data carried out to regenerate a symbol in the symbol data stream, and replace the particular symbol with the regenerated symbol.
- the crest factor of the symbol data stream can be reduced.
- significant perturbations in the values of the input vectors for the IFFT sufficient to cause the modified symbol to take a shape substantially different from the original one, can be accomplished by quite small and simple changes to the data being processed within the encoding system prior to the IFFT input stage.
- a symbol is detected whose final (time-domain) signal shape contains a peak higher than the threshold level, all or part of the processing of data which was performed in order to create that symbol is re-executed, this time making a change to some item of data which contributes to the symbol. If the resulting re-generated symbol has a lower peak value than the threshold, all is well and the revised symbol is sent.
- the modelling of subsequent processing is not used just to influence processing in the analogue front end, but also to reduce the crest factor.
- the effects of high crest factor are reduced both by reducing the crest factor and reducing the effects of the crest factor.
- the invention relates to apparatus for carrying out this method as well as to a computer program product for implementing the method.
- Figure 1 shows a schematic diagram of a first embodiment of an apparatus according to the invention.
- a digital data stream 10 is fed into a modulator 126 where it passes through a plurality of processing stages 12, 14.
- the output of these processing stages is a symbol data stream 26 including multi-tone symbols, and is stored in buffer 158.
- This symbol data stream is delivered from the buffer 158 to analogue front end 146 which contains a digital to analogue converter (DAC) 156 and a line driver, i.e. an amplifier 150.
- the line driver drives line 154, which may typically be a telephone line or other suitable interconnection or networking line.
- the symbol data stream 26 is also fed in the modulator 126 to a model 32.
- the model models the processing subsequently to be carried out in the analogue front end 146.
- the model 32 determines the peak amplitude the symbol will contain after passing through the analogue front end 146 and outputs a corresponding control signal 184 to the analogue front end 146.
- This control signal is used to control the analogue front end 146 in accordance with the derived symbol peak amplitude.
- analogue front end 146 only includes the DAC 156 and line driver 150.
- preprocessing 160 may be provided in the analogue front end 146, for example to oversample the incoming symbol data stream 26 to improve the digital to analogue conversion.
- Optional embodiments of the invention include regeneration control 28 to pass information back along signal path 161 to regenerate a symbol stored in the buffer 158 when the model 32 predicts too high a peak amplitude.
- the invention may also allow the control signal 184 to control the properties of other stages, such as AFE preprocessing stages, and/or of the digital to analogue converter.
- the line 154 is a phone line. Many other types of line are suitable for use in the present invention.
- the present invention may be applied to ADSL modems.
- the invention may be applied to alternative multi-tone signalling systems, such as COFDM (Carrierless Orthogonal Frequency Division Multiplex), a standard widely adopted for digital terrestrial TV broadcasting.
- COFDM Carrierless Orthogonal Frequency Division Multiplex
- a modulator 126 feeds data through a buffer 158 into the analogue front end 146 which contains a preprocessor 160, a DAC 156 and a line driver 150 to drive the line 154.
- the modulator 126 includes a model 32 and a control output 172.
- the purpose of the model 32 will be described later.
- the line driver 150 i.e. the final amplifier, is connected to a low voltage power supply 178 and to a high voltage power supply 180.
- a switch 182 switches the power supply to the line driver 150 between low voltage power supply 178 and high voltage power supply 180.
- the control signal 184 on control line 174 controls the switch 182 to normally use the low voltage power supply 178 but to switch the high voltage power supply 180 when a symbol amplitude peak on the symbol data stream is too high to be successfully or safely driven from the low voltage power supply 178.
- the power supply can be managed to optimise the power efficiency.
- the high efficiency, low voltage power supply 178 is normally used.
- the high voltage power supply 180 is used only when required since the amplifier consumes more power when running at the high voltage.
- the model 32 may include software to determine whether to switch the switch not merely on the peak amplitude in one symbol but that on adjacent symbols. This is because, at the margins where the high voltage or the low voltage power supply may be required, it may be the case that a sequence of symbols with peak amplitudes near the threshold level would require the use of the high voltage power supply, whereas an isolated symbol at that level could be successfully transmitted using the low voltage power supply. This will of course depend on the properties of the line driver 150 which may be readily determined by the skilled person.
- the line driver 150 may be implemented as one chip and a further chip 151 carries the preprocessing block 160 and digital to analogue converter 156.
- FIG 3 shows a flow diagram of the use of the ADSL modem of Figure 2 to transmit an input data streams of ATM cells 102.
- the cells are buffered 104, and idle cells 106 are inserted 108 as required.
- the cell payload is then scrambled 110, and a cyclic redundancy check 112 performed.
- the ATM cells are then combined by framing 114, adding fast bytes 116 where required. Scrambling 118 is then performed, followed optionally by Reed-Solomon Forward error correction 120.
- Another, interleaved data path 124 is also shown, having the same steps except that sync bytes 117 are used instead of fast bytes 116, and there is additionally a final step of convolutional interleaving 125.
- the cells of the two data paths 100, 124 can then be merged 122.
- the framed, merged and scrambled ATM cells are then passed to the modulator 126 which carries out the steps of tone ordering 128, optional trellis encoding 130, constellation encoding 132, gain scaling 134, and inverse Fourier transform 136 to produce a stream of symbols each encoding some part of the ATM cell stream.
- Cyclic prefixes are inserted (step 144).
- the symbols including prefixes are then passed to the symbol buffer 158 for buffering 142, and also to a modelling step 162, to be described below.
- cyclic prefix insertion 144 is carried out after symbol buffering 142, but including cyclic prefix insertion at an earlier stage avoids the need to include the insertion of cyclic prefixes into the modelling of the analogue front end.
- the output from the symbol buffer is passed to the analogue front end 146.
- This carries out preprocessing 164.
- the preprocessing includes signal filtering (optional) and oversampling; in oversampling the sample rate of the incoming stream is increased, typically by a factor of 2, 4 or 8 relative to the sample rate emerging from the buffer 158.
- the oversampling function includes a low-pass filter; it and any signal filtering functions inevitably cause some changes in the relative phase and amplitude of the individual tones of the symbol.
- the oversampling eases the subsequent signal processing, in particular the next step of the digital to analogue conversion 148.
- the analogue signal is used to drive the line (step 152) using line driver 150.
- the AFE preprocessing model 32 models 162 the effect on the symbol of the AFE preprocessing 164, including the oversampling and any filtering. Since the AFE preprocessing is generally carried out in the digital domain, the skilled person will not have any difficulty in modelling the preprocessing.
- the peak value of the modelled symbol can then be detected (step 138). If the peak is above a predetermined value then regeneration of the symbol is performed, under regeneration control 140. The ways in which this is done will be described in more detail later.
- the peak value is to be detected after any regeneration attempt or attempts have been completed.
- the power supply switch decision is accordingly based on the modelled peak amplitude in the final, possibly regenerated buffered symbol, not on any earlier generated version of that symbol. Since the threshold for power supply switching will normally be lower than the regeneration attempt threshold, regeneration will normally be attempted first, in any case where switching to the higher power supply would apparently be needed.
- the line driver 150 ( Figure 2) is arranged to run on two power supplies, a low voltage power supply 178 at 5V and a high voltage power supply 180 at 12V.
- a switch 182 normally supplies low voltage power from low voltage power supply 178, but can be switched to supply high voltage from high voltage power supply 180 instead. As the skilled person will appreciate, these voltages may vary depending on the application.
- a control input 170 is provided on the AFE 146 connected to control output 172 on the modulator by a control line 174.
- the modulator compares the peak amplitude in the symbol data stream with a predetermined threshold and if the power exceeds a predetermined high voltage threshold then the modulator outputs a control signal 184 through control output 172, control line 174 and control input 170 to control the switch 182 to supply the line driver 150 with the higher voltage power supply 180.
- the threshold is not necessary the same threshold as that used to trigger regeneration. Indeed, the threshold will normally be lower.
- the effect of the crest factor is reduced in two ways. Firstly, the low efficiency caused by the crest factor is improved by allowing the line driver 150 to operate in a low voltage, high efficiency mode for most of the time, only reverting to a high voltage, low efficiency mode when required. Also, the crest factor itself is reduced by regenerating symbols when the final input to the line driver would otherwise have a peak level greater than the desired maximum level.
- the approach allows the AFE module to be a separate module, as presently common.
- the AFE is not required to carry out either determination of the power nor symbol regeneration processing. If instead of using the approach of the embodiment the pre-processing were to be carried out in the modulator, oversampled data would have to be transferred from the modulator to the DAC 156 in the AFE. An increase in data transfer rate over the already high rate would typically cause more power to be consumed in the modulator output drive circuitry, increase local electrical noise, and generally make achieving system design goals more difficult.
- the preprocessing 160 and DAC 156 units of the AFE 146 are implemented in one chip 151 and the modulator 126 and buffer 158 in another 153; the line-driver 150 is a third separate device.
- the digital modulator 126 is built on a small geometry, more expensive process, so as to keep its size down and maximise digital processing speed.
- the preprocessing and DAC units of the AFE 146 are built on a larger-geometry, but cheaper and slower silicon process, which makes the design of analogue elements easier, and reduces the chip's cost.
- the modelling unit 162 in the modulator in this example, is chosen to be another instance of the preprocessing unit 160 in the AFE 146. However, because of the smaller geometry, it takes up less space. Since it is on the faster modulator chip 126, it can be clocked much faster.
- practical implementations of the invention may include multiple output channels. Since digital logic (including the preprocessing unit 160) in the AFE 146 is generally clocked more slowly than the modulator 126, and for other reasons, one physical copy of the preprocessing circuitry 160 is then used in the AFE for each output channel supported, rather than multiplexing the circuitry across different channels. However, in the modulator 126, the faster clock speed allows the modelling circuitry 162 to be time division multiplexed across multiple channels, thereby saving space. Thus, the overhead associated with the duplication of the preprocessing circuitry 160 as the modelling unit 162 may be less than would at first be thought.
- the preprocessing model 162 may in alternative embodiments be conveniently stored as data in a memory for controlling a central processor 186 of the modulator.
- the preprocessing model used can readily be adapted for different analogue front ends simply by changing the model in software.
- each functional block shown in the Figures 2 and 3 within the Data Path modules, the Modulator module and the Regenerator Control module, could in principle be implemented either by hardware or by software, or by some combination of the two.
- the AFE module normally uses hardware blocks for its functions.
- the single most valuable point at which data changes can readily be performed is in the "fast" and/or "sync" bytes which are defined to occur in ADSL data symbols.
- the fast and sync bytes are overhead bytes, not themselves part of the stream of data (usually an ATM cell stream) to be carried over the link, but associated with it, and physically carried as a part of the modulated signal.
- they contain control information used to manage the synchronization of data streams being transported over the ADSL link which were originated via a communication path whose control clock is asynchronous to the ADSL modem's own control clock.
- this capability is not required. Even where it is needed, it may actually be applied only rarely, leaving the byte available for application of the technique described here, most of the time.
- a fast or sync byte When a fast or sync byte is not carrying synchronization control values, it is defined to carry values of a fixed pattern, of the form XX0011X0 for the fast byte and XX0011XX for the sync byte.
- the bits shown as X can be freely set to either 0 or 1 as desired. With three or four bits whose value may be changed at will, there are a total of 8 or 16 possible combinations of 1s and 0s which may be created; therefore up to 7 or 15 attempts to re-generate a symbol are possible. This is more than adequate, in general.
- fast bytes occur at the start of many (e.g. 64 out of every 68) data frames of ADSL when "fast" (low latency) data streams are used. Changes in any X bit of the fast byte will cause much larger scale changes in the symbol because the stream scrambler is applied to this byte first in the fast data stream, so the scrambled form of almost all subsequent data bytes in the fast stream (but not the interleaved stream if also present) will in general be altered. Reed-Solomon encoding (if applied) also follows the scrambler so the added R-S parity bytes will in general take different values.
- the trellis encoder (if applied) will also cause changes to the output stream, because of any change in its input, over a given frame. Furthermore, since the fast byte is the first byte in each whole data frame (including also the interleaved data if present), this means that the trellis encoding of all or almost all tones in the symbol is liable to be modified by a single bit change in the fast byte.
- Sync bytes occur within most frames when interleaved streams are used.
- changes in any X bit of the normal sync byte pattern will affect all subsequent bytes in the interleaved portion of the data frame, through the application of the stream scrambler to the interleaved stream.
- the convolutional interleaver is applied, then older data from the interleaver's buffer, which will also appear in the final encoded symbol, will not be affected by the change; thus changes to the sync bytes are less effective.
- trellis encoding is in use, then all output data of the trellis encoder, starting from the interleaved part of the frame, will still be affected by a change in the sync byte.
- Both fast and sync bytes occur in many frames in a "dual latency" system where both fast and interleaved streams are active.
- either or both fast and sync bytes may be changed (a total of up to 7 X bits) to effect a significant change in the final time domain form of the encoded symbol.
- the fast or sync bytes are not available - 4 data frames (symbols) out of the 68 data frames in each ADSL "superframe" structure are defined not to carry them (the byte location in the frame is used for a different purpose in these frames), and in systems where synchronization must be performed, these bytes may occasionally carry values other than the default pattern with its three X bits.
- the fast and sync byte locations in the ADSL data frame are shared with use for other purposes, though these will generally be infrequently used; additionally one configuration reduces the number of frames carrying fast and /or sync bytes to 32 (rather than 64) out of every 68. If it is desired to modify the data and regenerate the symbol in these cases also, some other method must be found; alternative methods are given below.
- idle cells can provide a way to achieve the effect of the invention, without damage to real user data.
- Idle cells are ATM cells of a special reserved type, which are used to pad out a data stream. Idle cells are defined by a particular fixed pattern in the cell header - this is how they are recognized as idle cells when received. The payload of an idle cell is also defined as a fixed pattern, the same in every byte.
- ADSL the transmitting section of each modem is obliged to insert idle cells into the data stream whenever no user data cells are available to be transported. This is required because the physical data rate of a standard ADSL link is fixed at initialisation, and is maintained until the line is shut down or re-initialised. Since it is not possible to send "no data" when no user data is present, idle cells are sent instead, to maintain the flow. At the receiving end, idle cells are simply discarded - their contents are not related to real user data carried by the connection.
- the earliest available idle-cell payload byte in the data for the symbol should be so modified, since all modifications affect (by spreading) only the encoded form of later bytes in the stream (and hence, that portion of the stream which is carried in the rest of the symbol).
- ADSL connections only carry user data for a small proportion of the time, when considered on a long term basis; so idle cells will be very common in general. Since an idle cell does not contain any user data, its payload is not of interest, being fixed. Changing any bit in the payload of an idle cell will have no effect on the user data also carried by an ADSL connection.
- One proviso to this method is that one way of checking the error rate on an ADSL link, sometimes employed for purposes of link maintenance and management, is for the receiving modem to examine the payload bytes of idle cells before it discards them, comparing each byte against the fixed value it is defined to hold in any idle cell. Any errors found in the comparison are assumed to have arisen as a result of uncorrected errors in transmission of the data stream over the ADSL link. Some modems keep count of the error rate on this basis (measured as a moving average of the number of bits in idle cells which are found to be incorrect, divided by the total number of bits in the idle cells seen, over some measurement interval).
- One possible method is to define a limited set of modifications to idle cell payload bytes which can be attempted by the transmitting modem. Instead of the standard fixed payload byte value, a small number of alternative values (say, 3 out of the 255 remaining possibilities, or even just one value) could also be considered to be "legitimate" in idle cell payloads. In such a scheme, the receiving modem would be modified so as not to count such values in idle cell payloads as being errors, for purposes of error rate calculation.
- This modification still allows a high rate of true error detection, since the probability that a randomly corrupted idle cell payload byte takes one of 4 specific allowed values out of the 256 possible ones is only 1/64, or 1/256 if just one alternative value is allowed. Therefore with random, even error distribution, the true error rate and the measured one would differ by at most 6.25%, well within an appropriate level of accuracy in this context; furthermore in long term measurements it is possible to compensate for this difference.
Abstract
Description
Claims (9)
- A modulation method for multiple-tone signalling using a system with an analogue front end (146), comprising the steps of:feeding a symbol data stream of multiple tone symbols to a modelling means (32) and to a buffer (158) for onward transmission to the analogue front end (146);in the modelling means (32), modelling the peak amplitude that will be present in the symbol data stream after subsequent processing by the analogue front end (146);feeding forward a control signal based on the modelled peak amplitude from the modelling means (32) to the analogue front end (146);and outputting the symbol data stream from the buffer (158) through the analogue front end (146) under the control of the control signal, wherein the analogue front end (146) includes an amplifier (150) operable from a plurality of different voltage supply levels, and
- A method according to claim 1, including preprocessing the symbol data stream in the analogue front end (146), and modelling the preprocessing in the model (32).
- A method according to claim 2 wherein the modelling is carried out separately on each symbol.
- A method according to claim 3 further comprising processing an input data stream (10) through a plurality of intermediate processing stages (12, 14) and corresponding stages of intermediate data to generate the symbol data stream; and if the modelled peak amplitude in a particular symbol in the symbol data stream exceeds a predetermined threshold, amending predetermined intermediate data such that the input data is still represented by the intermediate data, carrying out the subsequent intermediate processing stages (12,14) on the intermediate data to regenerate the particular symbol in the symbol data stream, and replacing the particular symbol with the regenerated symbol.
- A multiple tone modem comprising:a modulator (126) for generating a symbol data stream of multiple tone symbols;a buffer (158) for buffering the symbol data stream for onward transmission to an analogue front end (146) for processing the symbol data stream, the analogue front end (146) including a digital to analogue converter (156) and a line driver (150) for driving a line;a modelling means (32) for processing the symbol data stream to predict the amplitude peaks that will be present in the symbol data stream after subsequent processing by the analogue front end (146) and for feeding forward a control signal based on the modelled amplitude peaks to the analogue front end (146);
- A multiple tone modem according to claim 5 wherein the analogue front end (146) further comprises a preprocessing module for preprocessing the symbol data stream, and wherein the model (32) models the preprocessing.
- A multiple tone modem according to claim 5 wherein the model (32) models the peak amplitude separately for each symbol in the symbol data stream.
- A multiple tone modem according to claim 7 wherein: the modulator (126) includes a plurality of intermediate processing stages (12, 14) for processing an input data stream through a plurality of stages of intermediate data and generating the symbol data stream; and the modulator further comprises a regeneration control system actuated if the modelled peak amplitude in a symbol exceeds a predetermined threshold to amend predetermined intermediate data such that the input data is still represented by the intermediate data, and to carry out the subsequent intermediate processing stages (12, 14) on the amended intermediate data to regenerate a replacement symbol.
- A multiple tone transmission system comprising:a transmitter includinga modulator (126) for generating a symbol data stream of multiple tone symbols;a buffer (158) for buffering the symbol data stream for onward transmission to an analogue front end (146) for processing the symbol data stream, the analogue front end (146) including a digital to analogue converter (156) and a line driver (150) for driving a line;and a modelling means (32) for processing the symbol data stream to predict the amplitude peaks that will be present in the symbol data stream after subsequent processing by the analogue front end (146) and for feeding forward a control signal based on the modelled amplitude peaks to the analogue front end (146);
the system further comprising a transmission line (154); and
a receiver connected to the transmission line (154) to decode the transmitted data stream.
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
US09/921,757 US7020188B2 (en) | 2001-08-06 | 2001-08-06 | Multi-tone transmission |
US921757 | 2001-08-06 |
Publications (2)
Publication Number | Publication Date |
---|---|
EP1286511A1 EP1286511A1 (en) | 2003-02-26 |
EP1286511B1 true EP1286511B1 (en) | 2005-01-26 |
Family
ID=25445933
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
EP02255456A Expired - Lifetime EP1286511B1 (en) | 2001-08-06 | 2002-08-05 | Peak power control in a Multi-tone transmission system |
Country Status (4)
Country | Link |
---|---|
US (1) | US7020188B2 (en) |
EP (1) | EP1286511B1 (en) |
AT (1) | ATE288165T1 (en) |
DE (1) | DE60202720T2 (en) |
Families Citing this family (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20040125869A1 (en) * | 2002-12-31 | 2004-07-01 | May Michael R. | Method and apparatus for non-intrusive transceiver property adjustment |
US7286605B2 (en) * | 2003-09-30 | 2007-10-23 | Infineon Technologies Ag | Method and apparatus for reducing a crest factor of a multi-tone data signal |
WO2006106503A2 (en) * | 2005-04-06 | 2006-10-12 | Siano Mobile Silicon Ltd. | A method for improving the performance of ofdm receiver and a receiver using the method |
US8681841B2 (en) * | 2009-11-09 | 2014-03-25 | Adeptence, Llc | Method and apparatus for a single-carrier wireless communication system |
US10742467B1 (en) * | 2019-07-10 | 2020-08-11 | United States Of America As Represented By Secretary Of The Navy | Digital dynamic delay for analog power savings in multicarrier burst waveforms |
Family Cites Families (12)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US4831637A (en) * | 1984-06-19 | 1989-05-16 | American Telephone And Telegraph Company | Apparatus and technique for timing jitter cancellation in a data receiver |
US4679227A (en) | 1985-05-20 | 1987-07-07 | Telebit Corporation | Ensemble modem structure for imperfect transmission media |
CA2054049C (en) * | 1990-11-05 | 1996-02-06 | Henry L. Kazecki | Apparatus and method for removing distortion in a received signal |
EP0719001A1 (en) | 1994-12-22 | 1996-06-26 | ALCATEL BELL Naamloze Vennootschap | DMT modulator |
JPH09153882A (en) * | 1995-09-25 | 1997-06-10 | Victor Co Of Japan Ltd | Orthogonal frequency division multiple signal transmission system, transmitting device and receiving device |
US5970093A (en) * | 1996-01-23 | 1999-10-19 | Tiernan Communications, Inc. | Fractionally-spaced adaptively-equalized self-recovering digital receiver for amplitude-Phase modulated signals |
US6028486A (en) | 1997-10-07 | 2000-02-22 | Telefonaktiebolaget Lm Ericsson | Method and apparatus for reducing power dissipation in multi-carrier amplifiers |
US6215354B1 (en) | 1998-03-06 | 2001-04-10 | Fujant, Inc. | Closed loop calibration for an amplitude reconstruction amplifier |
JP2003500932A (en) | 1999-05-21 | 2003-01-07 | コーニンクレッカ フィリップス エレクトロニクス エヌ ヴィ | Peak reduction for averaging power ratio in multicarrier transmission |
DE19927952A1 (en) * | 1999-06-18 | 2001-01-04 | Fraunhofer Ges Forschung | Device and method for predistorting a transmission signal to be transmitted over a non-linear transmission path |
GB9928184D0 (en) * | 1999-11-29 | 2000-01-26 | British Broadcasting Corp | Improvements in ofdm transmitters and recievers |
GB2365283B (en) * | 2000-07-21 | 2004-07-07 | British Broadcasting Corp | Many-carrier signal and transmission and reception thereof |
-
2001
- 2001-08-06 US US09/921,757 patent/US7020188B2/en not_active Expired - Fee Related
-
2002
- 2002-08-05 AT AT02255456T patent/ATE288165T1/en not_active IP Right Cessation
- 2002-08-05 DE DE60202720T patent/DE60202720T2/en not_active Expired - Lifetime
- 2002-08-05 EP EP02255456A patent/EP1286511B1/en not_active Expired - Lifetime
Also Published As
Publication number | Publication date |
---|---|
DE60202720T2 (en) | 2006-01-12 |
US7020188B2 (en) | 2006-03-28 |
DE60202720D1 (en) | 2005-03-03 |
EP1286511A1 (en) | 2003-02-26 |
ATE288165T1 (en) | 2005-02-15 |
US20030026331A1 (en) | 2003-02-06 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US7903757B2 (en) | Multi-tone transmission | |
US7403569B2 (en) | Efficient low-power mode for multicarrier communications | |
EP0923823B1 (en) | Improvements in, or relating to, multi-carrier transmission systems | |
AU696798B2 (en) | Frame synchronization in multicarrier transmission systems | |
US6628722B1 (en) | Decoding technique in discrete multi-tone (DMT) based communications systems | |
EP0922344B1 (en) | Improvements in, or relating to, multi-carrier transmission systems | |
US6954505B2 (en) | Discrete multitone modulation with reduced peak-to-average ratio using unloaded subchannels | |
US20010022810A1 (en) | Discrete multi-tone processor in asymmetric digital subscriber line modem | |
EP1283627B1 (en) | Multitone transmission with control of peak power | |
EP1286511B1 (en) | Peak power control in a Multi-tone transmission system | |
US6885699B2 (en) | Semi-stationary quiescent mode transmission | |
KR20070015449A (en) | Configurable dsl transceiver | |
US7177350B1 (en) | Receiver transparent quiescent mode operation in a DSL transmission system | |
Faraj et al. | A Review of Peak to Average Power Ratio Reduction Schemes in OFDM Systems | |
US6965637B2 (en) | System and method for controlling line driver power in a shared digital subscriber line modem | |
EP1269670A2 (en) | Transmission of stationary signals in quiescent mode | |
EP1248429B1 (en) | Discrete multitone modulation with reduced peak-to-average power ratio, using unloaded subchannels | |
KR100758765B1 (en) | Apparatus and method for reduction of peak-to-average power ration in DMT modem | |
SE506636C2 (en) | Receiver for use with multicarrier transmission system using orthogonal carriers | |
SE506644C2 (en) | Channel value calculating and equalisation system for use in multiple carrier wave system | |
SE506643C2 (en) | Multi-carrier wave transmission system e.g. for VDSL system | |
SE506637C2 (en) | Receiver for use with multicarrier transmission system using orthogonal carriers | |
SE506635C2 (en) | Receiver for multiple carrier wave transmission system | |
SE506634C2 (en) | Multiple carrier wave transmission system | |
SE506638C2 (en) | Receiver used with multicarrier transmission system using orthogonal carriers |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PUAI | Public reference made under article 153(3) epc to a published international application that has entered the european phase |
Free format text: ORIGINAL CODE: 0009012 |
|
AK | Designated contracting states |
Kind code of ref document: A1 Designated state(s): AT BE BG CH CY CZ DE DK EE ES FI FR GB GR IE IT LI LU MC NL PT SE SK TR |
|
AX | Request for extension of the european patent |
Extension state: AL LT LV MK RO SI |
|
17P | Request for examination filed |
Effective date: 20030822 |
|
AKX | Designation fees paid |
Designated state(s): AT BE BG CH CY CZ DE DK EE ES FI FR GB GR IE IT LI LU MC NL PT SE SK TR |
|
17Q | First examination report despatched |
Effective date: 20031021 |
|
GRAP | Despatch of communication of intention to grant a patent |
Free format text: ORIGINAL CODE: EPIDOSNIGR1 |
|
GRAS | Grant fee paid |
Free format text: ORIGINAL CODE: EPIDOSNIGR3 |
|
GRAA | (expected) grant |
Free format text: ORIGINAL CODE: 0009210 |
|
AK | Designated contracting states |
Kind code of ref document: B1 Designated state(s): AT BE BG CH CY CZ DE DK EE ES FI FR GB GR IE IT LI LU MC NL PT SE SK TR |
|
PG25 | Lapsed in a contracting state [announced via postgrant information from national office to epo] |
Ref country code: CZ Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT Effective date: 20050126 Ref country code: CH Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT Effective date: 20050126 Ref country code: BE Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT Effective date: 20050126 Ref country code: IT Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT;WARNING: LAPSES OF ITALIAN PATENTS WITH EFFECTIVE DATE BEFORE 2007 MAY HAVE OCCURRED AT ANY TIME BEFORE 2007. THE CORRECT EFFECTIVE DATE MAY BE DIFFERENT FROM THE ONE RECORDED. Effective date: 20050126 Ref country code: FI Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT Effective date: 20050126 Ref country code: BG Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT Effective date: 20050126 Ref country code: TR Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT Effective date: 20050126 Ref country code: AT Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT Effective date: 20050126 Ref country code: LI Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT Effective date: 20050126 Ref country code: EE Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT Effective date: 20050126 Ref country code: SK Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT Effective date: 20050126 Ref country code: NL Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT Effective date: 20050126 |
|
REG | Reference to a national code |
Ref country code: GB Ref legal event code: FG4D |
|
REG | Reference to a national code |
Ref country code: CH Ref legal event code: EP |
|
REG | Reference to a national code |
Ref country code: IE Ref legal event code: FG4D |
|
REF | Corresponds to: |
Ref document number: 60202720 Country of ref document: DE Date of ref document: 20050303 Kind code of ref document: P |
|
PG25 | Lapsed in a contracting state [announced via postgrant information from national office to epo] |
Ref country code: SE Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT Effective date: 20050426 Ref country code: DK Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT Effective date: 20050426 Ref country code: GR Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT Effective date: 20050426 |
|
PG25 | Lapsed in a contracting state [announced via postgrant information from national office to epo] |
Ref country code: ES Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT Effective date: 20050507 |
|
NLV1 | Nl: lapsed or annulled due to failure to fulfill the requirements of art. 29p and 29m of the patents act | ||
REG | Reference to a national code |
Ref country code: CH Ref legal event code: PL |
|
PG25 | Lapsed in a contracting state [announced via postgrant information from national office to epo] |
Ref country code: IE Free format text: LAPSE BECAUSE OF NON-PAYMENT OF DUE FEES Effective date: 20050805 Ref country code: CY Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT Effective date: 20050805 Ref country code: LU Free format text: LAPSE BECAUSE OF NON-PAYMENT OF DUE FEES Effective date: 20050805 |
|
PG25 | Lapsed in a contracting state [announced via postgrant information from national office to epo] |
Ref country code: MC Free format text: LAPSE BECAUSE OF NON-PAYMENT OF DUE FEES Effective date: 20050831 |
|
PLBE | No opposition filed within time limit |
Free format text: ORIGINAL CODE: 0009261 |
|
STAA | Information on the status of an ep patent application or granted ep patent |
Free format text: STATUS: NO OPPOSITION FILED WITHIN TIME LIMIT |
|
26N | No opposition filed |
Effective date: 20051027 |
|
ET | Fr: translation filed | ||
REG | Reference to a national code |
Ref country code: IE Ref legal event code: MM4A |
|
PG25 | Lapsed in a contracting state [announced via postgrant information from national office to epo] |
Ref country code: PT Free format text: LAPSE BECAUSE OF NON-PAYMENT OF DUE FEES Effective date: 20050626 |
|
PGFP | Annual fee paid to national office [announced via postgrant information from national office to epo] |
Ref country code: FR Payment date: 20130820 Year of fee payment: 12 Ref country code: GB Payment date: 20130823 Year of fee payment: 12 |
|
GBPC | Gb: european patent ceased through non-payment of renewal fee |
Effective date: 20140805 |
|
REG | Reference to a national code |
Ref country code: FR Ref legal event code: ST Effective date: 20150430 |
|
PG25 | Lapsed in a contracting state [announced via postgrant information from national office to epo] |
Ref country code: GB Free format text: LAPSE BECAUSE OF NON-PAYMENT OF DUE FEES Effective date: 20140805 |
|
PG25 | Lapsed in a contracting state [announced via postgrant information from national office to epo] |
Ref country code: FR Free format text: LAPSE BECAUSE OF NON-PAYMENT OF DUE FEES Effective date: 20140901 |
|
PGFP | Annual fee paid to national office [announced via postgrant information from national office to epo] |
Ref country code: DE Payment date: 20160831 Year of fee payment: 15 |
|
REG | Reference to a national code |
Ref country code: DE Ref legal event code: R082 Ref document number: 60202720 Country of ref document: DE Representative=s name: BOSCH JEHLE PATENTANWALTSGESELLSCHAFT MBH, DE Ref country code: DE Ref legal event code: R081 Ref document number: 60202720 Country of ref document: DE Owner name: AVAGO TECHNOLOGIES GENERAL IP (SINGAPORE) PTE., SG Free format text: FORMER OWNER: BROADCOM CORP., IRVINE, CALIF., US |
|
REG | Reference to a national code |
Ref country code: DE Ref legal event code: R119 Ref document number: 60202720 Country of ref document: DE |
|
PG25 | Lapsed in a contracting state [announced via postgrant information from national office to epo] |
Ref country code: DE Free format text: LAPSE BECAUSE OF NON-PAYMENT OF DUE FEES Effective date: 20180301 |